Re: [Discuss-gnuradio] use of gr-specest for direction finding of a 1X2 MIMO system

[Marcus Müller]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Chandan,

the specest (spectral estimation toolbox) uses MUSIC to estimate the
spectrum; it does that by using the MUSIC algorithm on an unmodified
autocorrelation matrix estimate of *one* input sample stream; if you
supply different data, MUSIC will estimate different things.
You'll find the algorithm in specest_music_armadillo_impl.cc; I
believe it's easy to understand if you've done MUSIC before.

As a side note: Benchmarks have shown that for small signal spaces (in
your case: few incident waves from different directions at once), the
ESPRIT algorithm (also in gr_specest!) performs slightly better, and
you get the "strongest" frequencies/directions in a vector, and don't
have to search the MUSIC pseudospectrum for local maxima.

Greetings,
Marcus


On 09.03.2014 12:12, Chandan Pradhan wrote:
> hello, I downloaded and build gr-specest block from github site " 
> https://github.com/kit-cel/gr-specest"; into gnu radio 3.6.5.1, But
> while make test  i got the following result:
> 
> Test project /home/testbed/Downloads/gr-specest-master/build Start
> 1: qa_specest_arburg_impl 1/15 Test  #1: qa_specest_arburg_impl
> ...................   Passed    0.54 sec Start  2:
> qa_specest_cyclo_fam_calcspectrum_impl 2/15 Test  #2:
> qa_specest_cyclo_fam_calcspectrum_impl ...   Passed    0.07 sec 
> Start  3: qa_specest_esprit_fortran_impl 3/15 Test  #3:
> qa_specest_esprit_fortran_impl ...........   Passed    0.24 sec 
> Start  4: qa_specest_burg 4/15 Test  #4: qa_specest_burg
> ..........................   Passed    1.95 sec Start  5:
> qa_specest_cyclo_fam 5/15 Test  #5: qa_specest_cyclo_fam
> .....................***Failed    0.55 sec Start  6:
> qa_specest_esprit 6/15 Test  #6: qa_specest_esprit
> ........................   Passed    0.31 sec Start  7:
> qa_specest_esprit_spectrum_vcf 7/15 Test  #7:
> qa_specest_esprit_spectrum_vcf ...........***Failed    0.30 sec 
> Start  8: qa_specest_esprit_vcf 8/15 Test  #8:
> qa_specest_esprit_vcf ....................   Passed    2.44 sec 
> Start  9: qa_specest_fcov 9/15 Test  #9: qa_specest_fcov
> ..........................   Passed    0.29 sec Start 10:
> qa_specest_fmcov 10/15 Test #10: qa_specest_fmcov
> .........................   Passed    0.29 sec Start 11:
> qa_specest_gendpss 11/15 Test #11: qa_specest_gendpss
> .......................   Passed    0.27 sec Start 12:
> qa_specest_mtm 12/15 Test #12: qa_specest_mtm
> ...........................   Passed    0.82 sec Start 13:
> qa_specest_vectormanip 13/15 Test #13: qa_specest_vectormanip
> ...................   Passed    0.29 sec Start 14:
> qa_specest_welch 14/15 Test #14: qa_specest_welch
> .........................***Failed    0.29 sec Start 15:
> qa_specest_welchsp 15/15 Test #15: qa_specest_welchsp
> .......................***Failed    0.33 sec
> 
> 73% tests passed, 4 tests failed out of 15
> 
> Total Test time (real) =   8.98 sec
> 
> The following tests FAILED: 5 - qa_specest_cyclo_fam (Failed) 7 -
> qa_specest_esprit_spectrum_vcf (Failed) 14 - qa_specest_welch
> (Failed) 15 - qa_specest_welchsp (Failed) Errors while running
> CTest make: *** [test] Error 8
> 
> 
> 
> futhur could i get any documentation which could teel me how to use
> root music avaliable gr-spect for the purpose of direction finding.
> currently i have one transmitter and two synchronized receives and
> i intend to find DOA of my transmitter.
> 
> 
> On Sun, Mar 9, 2014 at 1:37 PM, Moritz Fischer
> <address@hidden>wrote:
> 
>> 
>> On Mar 8, 2014 7:38 PM, "Chandan Pradhan"
>> <address@hidden> wrote:
>>> 
>>> can the  gr-specest be used to implement root music algorithm
>>> for
>> direction finding??
>> 
>> Yes.
>> 
> 
> 
> 
> _______________________________________________ Discuss-gnuradio
> mailing list address@hidden 
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJTHFUJAAoJEBQ6EdjyzlHtViwH/jwAb8+/4hh7bKfhy/jkKBJk
MGM2udvngvNNTaB/iaq3G4f1hPMe45x33W7NFW6vpXckb/R1+hRW+UcDjW4pBKzZ
VgVLoy7k1At2E63at+gWBz6e6dWmGWkfB5TXaqrOfWs7nU8S7RTchmMtxc5zfCQ6
VI13WUtmHgwyxdzfptOEoiqSUrP47D9u2be45pC+PWiv7cWGqTLip8Th6kS6eNdw
Wznz6Cm5kvSNkOl/qRRRouZ7OkAXZKeLLRg2spKKiKuyjkhLqKhz8Ix9XObHsSu4
ZjdqkuOPagD9Por9KjjPL+d208hxLbRkA7IczEiz8hOIwLiHGgVaAlwQ1qahFUs=
=1806
-----END PGP SIGNATURE-----